What an AI receptionist can and can't do for a clinic
Every few months a new product promises an AI that will "run your front desk." Clinic owners are right to be skeptical. A front desk isn't a script. It's a person reading a worried face, squeezing in a child with a toothache, and knowing which patient needs a gentler tone.
An AI assistant can still take real work off that desk. The trick is being clear about which work. Here is an honest map: what it does well, where it needs a person to approve, and what it should never be asked to do.
What it can do: real work on real records
The useful kind of AI assistant isn't a chatbot that answers questions about your opening hours. It works inside the same system your team uses, on the same patients, calendar, and invoices.
In Naab, the text assistant (on Professional and above) can:
- Book, move, and cancel visits.
- Find the next free slot with a particular dentist or doctor.
- Pull up today's list or a patient's file without anyone clicking through menus.
- Add a treatment to a patient's treatment plan, including where it goes on the chart.
- Create invoices and record payments.
- Send a WhatsApp message to a patient.
A receptionist can type "Book Sara on Thursday with Dr. Khaled" while another patient is standing at the desk, and the visit appears on the calendar. There's no magic in it. It means fewer clicks at the busiest moments of the day, and nobody having to remember where a setting lives.
Notice what that list has in common. Every action is something a staff member asked for, and every result lands somewhere the team can see and check. That's the right shape for AI in a clinic.
Help with WhatsApp replies
Most of a clinic's messages are variations on the same few questions. Is there anything earlier this week? How much do I still owe? Can I move my visit to the afternoon?
On Professional and above, the AI can help your team answer on WhatsApp. Because it can see the calendar and the patient's balance, an answer to "do you have anything Tuesday?" can start from the real free slots rather than a guess. The conversation stays in the clinic's shared inbox, so everyone on the team can read what was said.
A good rule of thumb: let the AI help with logistics, like times, balances, and directions, and keep a person on anything that isn't logistics.
When it answers the phone
The busiest moment at a clinic is usually the moment the phone rings. The receptionist is taking a payment, the assistant is in the treatment room, and the call goes unanswered. That patient may simply try the next clinic.
On the Enterprise plan, a clinic can choose to let the AI answer incoming WhatsApp calls. It doesn't confirm anything on its own. It does one of two things:
- It takes a message for the team.
- It takes a booking request: who called, what they need, and when they would like to come.
A booking request waits on the dashboard. A staff member opens it, sees a warning if the time clashes with another visit, and approves or declines it. Approving creates the visit, and can create a new patient if the caller wasn't on file yet. If someone on the team is free while the AI is on a call, they can stop it and take over.
That approval step is the point. The AI makes sure the call isn't lost. A person still decides what goes on the calendar.
What it shouldn't do
Being honest about the limits is what makes the rest trustworthy. Three kinds of work should stay with people.
Clinical judgement. Is this swelling urgent? Can I take a painkiller with my blood pressure pills? Should my child's loose tooth hurt this much? These questions need a dentist or doctor who knows the patient. An AI that answers them confidently is a risk, not a feature. The right move is to get the message to a clinician quickly.
Sensitive conversations. A complaint about a treatment that went badly. A family arguing over a bill. A patient who has just heard difficult news. These need a human voice, some patience, and often a decision about money or goodwill that only the owner can make.
Deciding without approval. Refunds, discounts, squeezing someone into a fully booked day, or changing a treatment plan are decisions, not tasks. The AI can carry them out when a staff member asks. It shouldn't be the one deciding.
None of this means the AI replaces your receptionist. It means your receptionist spends less time on clicks and more time on the patients in front of them.
How the AI is paid for
AI work in Naab is counted in AI actions. Paid plans include a monthly allowance, and larger plans include more. If a busy month runs past it, you can buy extra credits at any time, on any plan, including Free. You also choose what happens when the allowance runs out: stop, or keep going at a higher rate.
In practice, that means you can start small. Use the assistant for the tasks your team finds tedious, watch how much it uses, and adjust from there.
A sensible way to start
- Start with the calendar. Booking, moving, and finding free slots are where the help shows up first, and the results are easy to check.
- Write down what always goes to a clinician. Make sure everyone knows the list, including whoever reviews messages the AI took.
- Check booking requests at set times. First thing, after lunch, and before closing is a good rhythm.
- Review a week of conversations together. Keep what worked and tighten what didn't.
You can see how the assistant fits with the calendar and the inbox on the front desk page, and which plan includes what on the pricing page. The trial is full Enterprise with no card, so you can try call answering yourself before you decide.